Area
CLI
Problem or motivation
Add support for evaluating and judging files generated by an agent during an OPFOR evaluation. This enables testing whether an agent produces safe, correct, and policy-compliant artifacts when its output is a file rather than plain text.
Proposed solution
Extend agent execution to capture generated artifacts:
- Files created by the agent
- Filename and extension
- MIME type
- File size
- Hash/checksum
- Generation metadata
Alternatives considered
No response
Area
CLI
Problem or motivation
Add support for evaluating and judging files generated by an agent during an OPFOR evaluation. This enables testing whether an agent produces safe, correct, and policy-compliant artifacts when its output is a file rather than plain text.
Proposed solution
Extend agent execution to capture generated artifacts:
Alternatives considered
No response